Strong acid solutions have been widely used in acid traps
to determine concentrations of ammonia in ambient air or exhaust air stream. A literature survey indicates the method has a long history and a wide variation in use. Trough a series of studies, this paper examines several factors including volume of the acid, depth of the acid, and airow rate; that might aect the e ciency of sulfuric acid traps and recommends steps researchers and other users may take to ensure reliable results from this method. Te results from these series of studies indicate: (i) an inverse relationship between the e ciency of the acid traps and the amount of ammonia to be trapped even when the capacity of the acid trap is excessively greater than the maximum theoretical stoichiometric capacity needed to dissolve all of the ammonia, (ii) for the same volume of acid, the e ciency of the acid trap increased with the acid depth but overall, the e ciency at any given acid depth decreased as the amount of ammonia through the trap increased, and (iii) at the two airow rates examined in this study (0.5 and 1.0 L/ min) the e ciency of the acid traps decreased at similar rates as the concentration of ammonia in the sample air increased but the e ciency of the trap was signicantly higher at the lower airow rate. To obtain reliable measurements from this method, therefore, multi-point calibrations within the entire range of target measurements is recommended to provide accurate corrections of the measurements. Measuring Concentrations of Ammonia in Ambient Air or Exhaust Air Stream using Acid Traps P. M. Ndegwa* and V. K. Vaddella Washington State University A. N. Hristov Pennsylvania State University H. S. Joo Washington State University T determine ammonia in a stream of air or concentrations in ambient air, a sample of air is pulled through an acid solution that traps the ammonia. Most of the ammonia traps use either boric acid, sulfuric acid, or some other strong acid. Boric acid is chosen if titration is used to determine amount of ammonia trapped while sulfuric acid is used when colorimetric methods are used (APHA, 1998). Te choice of either method (titration or colorimetric) can also depend on the concentration of ammonia in solution. In general, colorimetric methods are used to determine lower concentrations of ammonia in solutions (down to 2 mg/L) than titration methods, although interferences may be a problem in some cases with respect to colorimetric methods. Ammonia absorption in the acid is a neutralization reaction be- hind the titration method. Tis understanding helps to estimate the amount of acid solution to trap most of the ammonia in the air sample. If this principle is not well understood or is not used to make these estimations, the amount of acid could be overwhelmed by the amount of ammonia in the air sample. Such situations results in poor trapping of the ammonia and hence underestimation of ammonia concentration in the air sample in question. Te other extreme is the use of excessive amount of the acid in the trap. Al- though this situation does not really amount to a signicant expense (because acid is inexpensive) the resulting ammonia concentration in the trap acid solution may be below detection limit. A review of the use of acid traps using dilute sulfuric acid solution indicates a wide variation and lack of consistency in the use of this method. Todd et al. (2006) describes an acid trap system to evalu- ate eect of crude protein in the diets of beef cattle (Bos taurus) on ammonia emissions from beef cattle facilities. Tis system was based on the earlier work of both Shi et al. (2001) and Cole et al. (2005) in which both had used 100 mL of 0.9 mol/L sulfuric acid traps at airow rates of 3 L/min. Cole et al. (2005) study was also for evaluat- ing inuence of dietary crude protein on ammonia emissions from beef cattle facilities. On the other hand, Shi et al. (2001) had used their system to evaluate the e cacy of surface amendments in mitiga- tion of ammonia emission from beef cattle feedlots. Tere is neither indication of pre-estimation of the amount of acid necessary based on estimated ammonia in the air stream nor attempt to calibrate (i.e., determine trapping e ciencies) in either of these systems. P.M. Ndegwa, V. Vaddella, and H.S. Joo, Biological Systems Engineering, Washington State Univ., P.O. Box 646120, Pullman, WA 99164; A.N. Hristov, Dairy and Animal Science Dep., Pennsylvania State Univ., University Park, PA 16802. Copyright 2009 by the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including pho- tocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Published in J. Environ. Qual. 38:647653 (2009). doi:10.2134/jeq2008.0211 Received May 1 2008. *Corresponding author (ndegwa@wsu.edu). ASA, CSSA, SSSA 677 S. Segoe Rd., Madison, WI 53711 USA TECHNICAL REPORTS: ATMOSPHERIC POLLUTANTS AND TRACE GASES 648 Journal of Environmental Quality Volume 38 MarchApril 2009 Misselbrook et al. (2005a) used 75 mL of 0.02 mol/L or- thophosphoric acid with an airow rate of 4 L/min to evaluate eect of dietary manipulation in dairy cattle. Misselbrook et al. (2005a) conducted a one point calibration and reported a mean recovery of 97% for this single point calibration. Cabrera et al. (2001) describes another similar system used for deter- mining ammonia loss from soil surfaces that used 150 to 650 mL of 0.05 mol/L sulfuric acid. Similarly, this study also nei- ther indicates any attempt to rationalize the volume of the acid used nor calibration over a range of ammonia concentrations. Guiziou and Beline (2005) reports using 0.1 mol/L sulfuric acid traps to determine concentrations of ammonia in a stream of ventilation air from broiler houses. Details of the rate of air sampling and the volume of the acid used were not given. Rana and Mastrorilli (1998) used 250 mL 0.1 mol/L sulfuric at 4 L/ min to determine concentration of ammonia in air after eld application of manure. Again, no details of calibration were provided. In a similar work, Genermont et al. (1998) used a 0.075 mol/L sulfuric acid trap at an airow rate of 3 L/min to measure ammonia uxes after slurry spreading in the eld. And yet, in another similar study, McInnes et al. (1985) used 18 mL of 0.05 mol/L sulfuric acid at airow rates of 3.3 L/min to measure concentrations of ammonia in ambient air to estimate ammonia ux after application of fertilizer. More rigorous tests on ammonia trapping in acid solutions for measuring ammonia emissions from land applied ma- nure and manure storages were reported by Misselbrook et al. (2005b) and Xue et al. (1998), respectively. In the former study, two traps in series each containing 30 mL orthophosphoric acid solution were tested at three dierent concentrations (0.1, 0.01, and 0.001 mol/L) and at airow rates of 2 and 4 L/min. Te conclusion was that one ask was adequate while the airow rate and the acid strength had no signicant eects on the ammo- nia trapping e ciency. In the latter study, Xue et al. (1998) also used two traps in series each containing 1 mol/L sulfuric acid solution. Tey reported ammonia trapping e ciencies of 96.45 to 99.5% in the rst trap and 0.5 to 3.45% in the second trap. Tis system thus achieved nearly 100% trapping e ciency. Te airow rate in this system was not given. In a review paper on methods for ammonia concentration in air, Phillips et al. (2001) notes the acid trap method is sim- ple, cheap, and reliable and suitable for low concentrations of ammonia in air but requires a large excess dilute solution of a strong acid. Tis is rather subjective because there is no guid- ance on what excess acid solution means. In summary, there seems to be no consensus not only on the need to calibrate but also how to calibrate such a system. In addition, there seems to be lack of guidance in deciding what volume of the acid or airow rate to use for accurate results. Tis paper examines several factors that might aect the per- formance of sulfuric acid traps and recommends steps researchers may take to ensure reliable results from this method. Specically, this study examined e ciency of an acid trap: (i) over a range of ammonia concentration in an air stream, (ii) at two dierent amounts of acid, (iii) at two dierent acid depths, and (iv) at two dierent airow rates through the acid trap system. Methods and Materials System Confguration An ammonia trap system similar to those in the literature (Shi et al., 2001; Portejoie et al., 2004; Todd et al., 2006; Cole et al., 2005) was used in this study (Fig. 1). Tis system essen- tially consisted of a simulated-manure-storage or a simulated- manure-application system, an acid bottle to trap the emitted ammonia, a critical orice and a variable area ow-meter (Cat- alogue no. C-32460-42, Cole-Parmer Instrument Company; 5% full-scale accuracy) to regulate ow rate of the sweep-air, and a vacuum pump (Model DOA-P104-AA, Gast MFG Corp.) to pull air through the system. A 4-L completely sealed plastic container was used for the manure storage allowing only the entry and exit of headspace sweep-air. Tis air carry- ing emitted ammonia was delivered to the acid traps via equal lengths of 6.35 mm ID Teon tubing in all acid trap systems to cancel eects of ammonia adsorption on the tubing. Acid bottles were made of graduated Polypropylene cylinders sealed with vinyl stoppers with two short 6.35 mm OD glass tubes. To one end of one of these tubes was connected the ammonia carrier-air tubing while the other end was connected to a short 6.35 mm ID Teon tubing that delivered carrier-air to the bot- tom of the trap-bottle at the open end. Te other short 6.35 mm OD glass tube was connected to the vacuum end with a Teon tubing to exhaust the ammonia-scrubbed air out. To avoid the eect of other extraneous factors during a certain test condition, a pair of six parallel ammonia genera- tion and trapping systems were run concurrently. Tese sys- tems generated and trapped dierent ammonia concentrations staggered within a range of ammonia emissions between 50 and 300mg/d. All 12 trap systems were connected to a central sealed container which was in turn connected to a single vacu- um pump. Te vacuum created in this sealed container pulled the sweep-air through the manure holding containers and into the acid traps. Unless otherwise stated, all experimental runs were conducted over a period of 24 h. Critical orices and variable area ow-meters regulated the sweep-air ow through each acid trap system. Critical orices were made from 1-mL pipettes with tips cut to appropriate sizes to meter the desirable airow. Variable area ow-meters calibrated against a NIST traceable air ow-meter were used to monitor the air ows. Study I: Efect of Ammonia Flux on Trap Ef ciency Ammonium chloride solutions were prepared to concentra- tions ranging between 200 and 2300 mg NH 4 + /L to simulate ammonia emissions from manure storage or manure applica- tion in the eld ranging between 50 and 300 mg/d. To achieve this, 1.0 L/min ambient air was allowed to sweep over the headspace of the emitting surface after injection of 60 mL of 1 mol/L sodium hydroxide (NaOH) solution. Te NaOH solu- tion was added to the ammonium chloride solution to raise the pH of the solution and thus change ammonium (NH 4 + ) ions in solution to unionized volatile ammonia (NH 3 ), whose volatil- ization would be enhanced by the sweep-air in the headspace. Ndegwa et al.: Measuring Concentrations of Ammonia in Air 649 Te following six concentrations of the solution were used in this study: 200, 600, 1000, 1400, 1800, and 2300 mg NH 4 + /L to provide six dierent ammonia uxes. When ammonia is passed through the acid trapping solu- tion, it neutralizes some of the acid. Adequate acid stoichio- metric capacity of the trap system is thus critical to ensure all the ammonia in the sample air is absorbed without complete acid neutralization. For example, when sulfuric acid solution is used in the trap, stoichiometrically one mole of the acid can completely trap two moles of ammonia as shown in Eq. [1]. 2 4 3 4 2 4 H SO +2NH (NH ) SO [1] To evaluate the e ciency of the acid trap in a range of am- monia emissions within the acid solution capacity, 150 mL of 0.2 mol/L H 2 SO 4 (equivalent to 0.03 moles of acid) was used in the trap. Te volume of the acid was calculated using Eq. [1] based on the amount of acid necessary to trap 600 mg, which was ap- proximately double the estimated 300 mg (0.0167 moles) maxi- mum amount of ammonia emissions expected from the system. Arithmetically, this volume of acid would thus be more than ad- equate to dissolve all the ammonia emitted from the entire target emission range (50300 mg) as well as provide adequate safety capacity (~100%) to trap any over the limit emissions beyond the estimated 300 mg upper limit and up to 600 mg. Studies II: Efects of Volume In subsequent studies to evaluate the eects of the volume of the acid in the trap system, the system was upgraded to hold more than three times (approximately 470 mL) the volume of the acid over the basic system (study I) that held only 150 mL while maintaining the same depth of the acid. A bigger diame- ter acid bottle was used to achieve this. Te ammonia emission sources were also congured as in study I to generate similar ammonia emissions ranging between 50 and 300 mg over a 24-h period at an airow rate of 1 L/min. Study III: Efect of Depth Two sets each of six acid traps systems were run concur- rently to determine the eect of depth of the acid on e ciency of the acid trap. Each of the acid trap in both sets contained 900 mL of the acid solution. However, the acid bottles in one set had a bigger diameter than the acid bottles in the other set resulting into two dierent depths of the acid. Te acid bottles of the rst set of traps held the 900 mL of acid to a depth of 110 mm while the acid bottles of the second set of traps held the same acid volume to a depth of 310 mm. Te ammonia emission sources were also congured as in study I to generate similar ammonia emissions ranging between 50 and 300 mg over a 24-h period at an airow rate of 1 L/min. Study IV: Efect of Air Flow-Rate To evaluate the eect of airow rate through the acid trap, two airow rates of 0.5 and 1.0 L/min commonly used in simi- lar systems were selected to carry emitted ammonia. Similarly, each air ow rate was evaluated at six levels of ammonia uxes using a set of six acid trap systems. Te volumes of the acid in all the acid bottles in this experiment were maintained at 900 mL and to a depth of 310 mm. Te ammonia emission ux range was set up just as in previous experiments in this study as was the 24-h experimental runs. Te acid trap at 0.5 L/min was, however, run for 48 h to obtain a similar range of ammo- nia emission ux as in the 1.0 L/min acid-trap systems. Sample Analysis In all the experiments, samples of ammonium chloride solu- tion were taken before and after each experimental run to deter- mine the amount of ammonia emitted from the solution. Te amounts of ammonia remaining at the end of each experimental run were adjusted for the original addition of NaOH. A sample of the acid-bottle content was also taken at the end of each run to determine the amount of ammonia trapped. Before taking the samples from acid bottles and the ammonia chloride solu- tion storages, the contents were made to the original volumes using distilled water to account for lost water during each test. To avoid loss of ammonia between sampling and analysis, 1 mL of sulfuric acid was added to 9 mL of the samples to acidify the sample. Te ammonia concentrations in these samples were ana- lyzed according to standard methods (APHA, 1992). Analyses were done immediately (within 4 h) after sampling taking into account the volume changes from the acidication processes. Data Analysis Te e ciencies of the traps were calculated as a percentage of the ratio of the amount of ammonia captured in the trap to Fig. 1. Schematic of one unit of the acid trap system used in our study. 650 Journal of Environmental Quality Volume 38 MarchApril 2009 the amount of ammonia that was actually emitted from the ammonia source. For each factor, linear regressions were per- formed on respective e ciencies against amounts of ammonia emitted. To determine the similarities or dierences due to a given factor, an analysis of covariance (ANCOVA) was then conducted on the slopes and intercepts of the respective pair of linear regressions using SAS (SAS Institute, 2003). Te ANCOVA was rst conducted to compare the slopes of such a pair of linear regressions. If slopes were signicantly dierent, then no further analysis was conducted because this information was su cient to conclude that the linear regressions were signicantly dierent. However, if slopes were not signi- cantly dierent, then another ANCOVA was conducted to de- termine if the intercepts were either not signicantly dierent (indicating overlapping or nearly overlapping regression lines) or signicantly dierent (indicating parallel regression lines). Un- less otherwise stated, statistical analyses and comparisons were conducted at the signicant level of = 0.05 in all cases. Results and Discussion Studies I and II: Acid Trap Ef ciency and Efect of Volume Te variations of the acid trap e ciency with ammonia ux and volume of acid in the trap are shown in Fig. 2 and 3. Ammonia uxes were determined by the dierences in the concentrations of ammonia in the simulated storage before and after each experimental run. Te relationships between the e ciency of the trap and the ux of ammonia as well as the volume of the acid in the trap are evident in Fig. 2. In the 150 mL of the acid trap system only 50% of the stoichiometric capacity was used to dissolve the maximum expected ammo- nia emission during a period of 24 h. Te 470 mL acid trap system on the other hand had over six times the stoichiometric capacity of capturing the maximum expected ammonia release during a similar period. In particular, the trapping e ciency of the smaller volume (150 mL) acid trap was more erratic than that of the bigger volume (470 mL) acid trap. Te evaluation of the smaller acid trap was, therefore, performed twice at ap- proximately the same conditions and the two sets of data were used in subsequent analyses. In general, the e ciency of the trap evidently decreased as the ux of ammonia increased ir- respective of the volume of the acid in the trap. From the ANCOVA results, the slopes of the two linear regressions lines (Fig. 2) were not signicantly dierent (P value = 0.93; denoted by P hereafter) indicating e ciencies of the two acid traps decreased at about the same rate as the ux of ammonia increased. Te implication of this observation with respect to a one point calibration that is fairly common in the literature is that, if the calibration point falls within the measurement range then any correction made on the mea- sured data using this calibration point will tend to overestimate emissions below the calibration emission while underestimat- ing those measurements above this point. Tis kind of error in the correction of the measured data may end up reducing the dierences in the performance among treatments, which may result in wrong conclusions or recommendations. Te intercepts of the linear regression lines of the e ciency against the ux of ammonia at the two acid volumes (Fig. 2) also were not statistically dierent (P = 0.42) indicating that even the absolute e ciency of the trap did not signicantly increase with the threefold increase in the amount of the acid solution in the trap-bottle. Te results of these studies thus do not provide any evidence to justify use of more than double the required stoichio- metric capacity of the acid in the trap to enhance e ciency of the trap. Tis inference is further supported by the calibrations of the acid trap at both of the two volumes of the acid (Fig. 3). Te lin- ear regressions lines of the multi-point calibrations indicate good ts to the measured data, which is manifested in relatively high coe cients of determination of 0.967 for the 150 mL acid trap and 0.985 for the 470 mL acid trap. Multi-point calibration of the acid trap system following this approach thus provides a more accurate correction of the actual measurements: (i) against the ex- pected inability of the acid trap to trap all the ammonia, and (ii) against the change in the trapping ability due to the uctuation of the ammonia ux. Te latter factor is denitely not addressed by single-point calibration procedures although it is seems critical for accurate calibration of acid trap systems. In summary, to improve the reliability of this method, be- sides ensuring the acid trap has adequate stoichiometric capac- ity to trap the maximum estimated ammonia in the air stream, a multi-point calibration of the acid trap system within the estimated range of actual measurements is essential. In addi- tion, this calibration procedure is still critical for the reliability of this method even if the stoichiometric capacity of the acid was excessively beyond the maximum estimated concentration of ammonia in the air stream because the e ciency of the trap seems to decrease with increase in the concentrations of the ammonia in the air sample irrespective of the amount of the acid used. Study III: Efect of Depth Te eect of the depth of the acid on the performance of the acid trap is presented in Fig. 4. Te two traps used for this study both contained 900 mL at depths of 110 and 310 mm in their respective acid-bottles. Again, the e ciency of the trap in either case decreased with increase in the ammonia concen- tration in the air irrespective of the depth of the acid in the trap-bottle. Te slopes of the two regressions lines were not signicantly dierent (P = 0.91) demonstrating the e ciencies of the acid trap at both the two depths decreased at the same rates with increase of the ammonia ux. Te intercepts of the two respective regression lines (at the two acid depths) on the other hand were signicantly dierent (P = 0.04) indicating the e ciency of the trap was higher at the greater depth of the acid. Tere was approximately 8% average increase in the trap-e ciency achieved by using a smaller diameter acid-bottle that increased the depth of the acid from 110 to 310 mm. Te multi- point linear calibrations of the two traps shown in Fig. 5 in addition also indicate good ts to the measured vs. the predicted data as indi- cated by the high coe cients of determinations of 0.945 at a depth of 110 mm and 0.981 at a depth of 310 mm. Te two acid traps thus calibrated very well and as long as either is calibrated over the Ndegwa et al.: Measuring Concentrations of Ammonia in Air 651 measurements range, the corrected measurements should be reliable irrespective of the depth of acid in the trap bottle. Based on these results, we inferred that the depth of the acid is indeed an important factor to consider when setting up a trap and given a choice between a smaller diameter (more depth) and bigger diameter (less depth) acid-bottle for the same acid amount, a smaller diameter acid-bottle improves the absolute e ciency of the trap. Tis observation can intuitively be arrived at because the residence time of the air in the trap increases with the increase in depth of the acid in the trap-bottle. Te increased residence Fig. 2. Ef ciency of the ammonia trap as a function of ammonia fux at two diferent acid volumes (150 and 470 mL) maintained at a depth of 110 mm. Fig. 3. Prediction of actual ammonia emitted based on ammonia trapped in the 150- and 470-mL acid trap systems both maintained at 110 mm acid depth. Fig. 4. Ef ciency of the 900-mL acid trap as a function of ammonia fux at two diferent depths (110 and 310 mm). 652 Journal of Environmental Quality Volume 38 MarchApril 2009 time provides more time for the ammonia to be absorbed in the acid solution. However, the increased depth achieved from using a smaller diameter acid-bottle in the trap did not translate into more reliability of the trap system as clearly illustrated by the multi-point calibrations (Fig. 5) at either of the acid depths. Study IV: Efect of Airfow Rate Te eect of airow rate was studied at two airow rates (0.5 and 1.0 L/min) and the results are presented in Fig. 6 and 7. In general, the e ciency of the acid traps at both airow rates decreased with increase in the ammonia ux (Fig. 6). Te major dierence was the absolute amounts of ammonia carried from the ammonia source at these two airow rates. At an airow rate of 0.5 L/min (meaning lower air velocity sweep through the ammonia source) the experi- ment had to be run for 48 h to achieve the same range of ammonia emissions as at the airow rate of 1.0 L/min in 24 h. Statistically, the slopes of the two regression lines of e ciency against ammonia ux at the two airow rates (Fig. 6) were not sig- nicantly dierent (P = 0.90). We thus inferred that the e ciency of the acid trap decreased with increase in the concentrations of the ammonia in the air at either of the two airow rates and the rates of decrease of the e ciencies were not signicantly dier- ent. However, the intercepts were signicantly dierent (P value = 0.02) indicating some signicant dierences in the absolute e - ciency of the acid trap at the two dierent airow rates. Tis result is not in agreement with that of earlier work of Misselbrook et al. (2005b). Tese researchers studied the eects of two airow rates (2 and 4 L/min) on the e ciency of two acid-traps in series each containing 30 mL orthophosphoric acid solution at three dierent concentrations (0.1, 0.01, and 0.001 mol/L). Teir conclusions were that the airow rate and the acid strength had no signicant eect on the e ciency of the trap. In this study, lowering the air- ow rate from 1.0 to 0.5 L/min increased the e ciency of the trap by an average of 4%. Te results of our studies, however, seem to be logical from theoretical standpoint. Lower air-ow rates may translate into more residence time of the carrier-air in the trapping solution and thus improve the trapping e ciency. Te multi-point calibrations of the acid traps at both airow rates are presented in Fig. 7. Again, the acid traps linear regressions for the multi-point calibrations lines are statistically good ts to the measured and predicted data as indicated by the high coe cients of determination of 0.980 at the 1.0 L/min and 0.967 at 0.5 L/min. As in the previous two factors (amount and depth of the acid in the traps), airow rates do not seem to signicantly aect the multi- point calibrations even though in absolute terms, the lower airow rate was slightly more e cient at trapping the ammonia. Conclusions Te e ciency of the traps decreases with the amount of am- monia to be trapped even when the stoichiometeric capacity of the acid trap greatly exceeds the maximum required acid ca- pacity. Te absolute e ciency of the trap did not signicantly increase with the threefold increase in the amount of the acid solution in the trap-bottle maintained at the same depth. Tis observation thus did not provide any evidence to justify use of more than double the required stoichiometric capacity of the acid in the trap to enhance e ciency of the trap. For the same acid amount, the e ciency of the acid trap was higher at the larger depth compared to the smaller depth. However, the rates of decrease of the e ciencies with increasing ammonia ux at dierent acid depths were statistically not dif- ferent. When the volume of the acid provides adequate stoichio- metric capacity for trapping all the ammonia estimated in a given air sample, the depth of acid in the trap should be maximized. Increasing the depth of the acid by about three times (from 110310 mm) using a smaller diameter acid-bottle increased the absolute e ciency by an average of 8% in this study. Te e ciency of the acid trap decreased at the same rate with the increase in the ammonia ux at the two dierent airow rates (0.5 and 1.0 L/min in this study) through the trap system. How- ever, e ciency of the trap signicantly increased with decrease in the airow rate. Lowering the airow rate twofolds from 1.0 to 0.5 L/min increased the e ciency of the trap by average of 4%. For the acid trap method to provide the most reliable results of ammonia ux, a multi-point calibration of the acid trap sys- tem covering the entire actual measurement range is necessary to provide accurate correction factors for the actual measure- ments. Te results of this study collectively suggest that un- less the range of measurement is known beforehand the system should be calibrated after the actual measurements. Tis proce- Fig. 5. Prediction of ammonia emitted from trapped ammonia in a 900-mL acid trap at two diferent depths (110 and 330 mm). Ndegwa et al.: Measuring Concentrations of Ammonia in Air 653 dure ensures that the system is calibrated throughout the entire range of the actual measurements. Acknowledgments Tis was research was supported by funds from Agricultural Research Center, Washington State Univ., and the United Dairymen of Idaho. References APHA. 1992. Standard methods for the examination of water and wastewater. 18th ed. American Public Health Assoc., NW, Washington, DC. APHA. 1998. Standard methods for the examination of water and wastewater. 20th ed. American Public Health Assoc., NW, Washington, DC. Cabrera, M.L., D.E. Kissel, R.C. Davis, N.P. Qafoku, and W.I. Segars. 2001. Design and ammonia-recovery evaluation of a wind speed-sensitive chamber system. Soil Sci. Soc. Am. J. 65:13021306. Cole, N.A., R.N. Clark, R.W. Todd, C.R. Richardson, A. Gueye, L.W. Greene, and K. McBride. 2005. Inuence of dietary crude protein concentration and source on potential ammonia emissions from beef cattle manure. J. Anim. Sci. 83:722731. Genermont, S., P. Cellier, D. Flura, T. Morvant, and P. Laville. 1998. Measuring ammonia uxes after slurry spreading under actual eld conditions. Atmos. Environ. 32:279284. Guiziou, F., and F. Beline. 2005. In situ measurement of ammonia and greenhouse gas emissions from broiler houses in France. Bioresour. Technol. 96:203207. McInnes, K.J., D.E. Kissel, and E.T. Kanemasu. 1985. Estimating ammonia ux: A comparison between the integrated horizontal ux method and theoretical solutions of the diusion prole. Agron. J. 77:884889. Misselbrook, T.H., F.A. Nicholson, B.J. Chambers, and R.A. Johnson. 2005b. Measuring ammonia emissions from land applied manure: An intercomparison of commonly used samplers and techniques. Environ. Pollut. 135:389397. Misselbrook, T.H., J.M. Powell, G.A. Broderick, and J.H. Grabber. 2005a. Dietary manipulation in dairy cattle: Laboratory experiments to assess the inuence of ammonia emissions. J. Dairy Sci. 88:17651777. Phillips, V.R., D.S. Lee, R. Scholtens, J.A. Garland, and R.W. Sneath. 2001. A review of methods for measuring emission rates of ammonia from livestock buildings and slurry or manure stores, Part 2: Monitoring ux rates concentrations and airow rates. J. Agric. Eng. Res. 78:114. Portejoie, S., J.Y. Dourmad, J. Martinez, and Y. Lebreton. 2004. Eect of lowering dietary crude protein on nitrogen excretion, manure composition, and ammonia emission from fattening pigs. Livest. Prod. Sci. 91:4555. Rana, G., and M. Mastrorilli. 1998. Ammonia emissions from elds treated with green manure in a Mediterranean climate. Agric. For. Meteorol. 90:265274. SAS Institute. 2003. SAS Proprietary Software Release 9.1 (TS1MO). SAS Institute, Cary, NC. Shi, Y., D.B. Parker, N.A. Cole, B.W. Auvermann, and J.E. Mehlhorn. 2001. Surface amendments to minimize ammonia emissions from beef cattle feedlots. Trans. ASAE 44:677682. Todd, R.W., N.A. Cole, and R.N. Clark. 2006. Reducing crude protein in beef cattle diet reduces ammonia emissions from articial feedyard surfaces. J. Environ. Qual. 35:404410. Xue, S.K., S. Chen, and R.E. Hermanson. 1998. Measuring ammonia and hydrogen sulde emitted from manure storages. Trans. ASAE 41:11251130. Fig. 6. Ef ciency of the 900-mL acid trap system as a function of ammonia fux at two airfow rates (0.5 and 1.0 L/min) through the ammonia trap. Fig. 7. Prediction of actual ammonia emitted from trapped ammonia trapped at two diferent airfow rates (0.5 and 1.0 L/min) in a 900-mL acid trap system.